Muni Metro is still broken.
And new trains won’t solve the problem (but they’ll help.)
During the month of August, there were — I kid you not — exactly three days without substantial delays paralyzing San Francisco’s primary rail network. And that’s only when you consider the SFMTA’s definition of a delay. While most major subway systems would consider a half dozen trains stopped in a tunnel problematic, to Muni, that’s just a bit of congestion. “No delay.”
This mentality represents the issues that plague Muni today. Every single time a subway stops in a tunnel, it should be a problem.
How is it that, in 2016, cities around the world have offered efficient, reliable, high-speed subways for decades, while Muni still hasn’t quite figured out how to run a metro?
Why our system remains broken
After a decade of frustration, in the mid-1990s, we were promised that our Muni Metro woes would be solved by a transition from the “archaic,” “mis-designed,” “obsolete,” trains at the time (Boeing LRVs), to a sleek new fleet of Italian trains, billed as “state of the art.” Willie Brown exclaimed that finally there was “light at the end of the tunnel.”
“I have a feeling when I get on one of the new cars it won’t break down.”
— Joe Ginoli, 2002.
That feeling turned out to not be quite right, as the new trains, despite being sold for a 30-year lifespan, required Muni to spent $56 million to “completely rehabilitate” that very same fleet just 7 years later, citing a myriad of issues. Only three years after that, Muni decided it was time for a new fleet.
In 2010, Muni’s proposed solution to congestion underground was so-called double-berthing. After countless delays over six years, multiple false starts, “software bugs,” and tens of millions of dollars, the much-hyped double-berthing finally launched this year. There’s been virtually no improvement in congestion underground.
We’ve twice had major overhauls of the Automated Train Control System (which manages the trains in the subway) and a full replacement of the overhead wires.
Meanwhile, above ground, crowding got so bad five years ago on the N-Judah that people would routinely wait 20–30 minutes just to board a train, a problem that actually continues to plague the N today. Rather than increasing service to match demand, Muni introduced a supplemental bus, the NX.
Apparently, for Muni, the future of trains that can’t keep up is… a bus.
The N-Judah routinely switches back at 19th Avenue to try to mitigate demand, something the Grand Jury found to be a “callous disregard for the welfare of riders,” and a practice Supervisor Tang directly campaigned against. The T-Third regularly switches trains back early for the same reasons. Just this week, Muni introduced a new N-Judah switchback at Hillway to try and increase capacity East of the Sunset, though it‘s more of a temporary stop-gap than a real solution.
Whether building the Embarcadero turnaround, buying a new fleet of sleek Italian trains, a major upgrade to the automated train control system, adding double berthing, performing yet another upgrade to the automated train control system, supplemental bus service, managed switchbacks, or (yet another) fleet of new trains, there seems to be an infinite number of ways to overpromise and under-deliver, without ever addressing the roots of the problem.
We’re continually strung along with promises that haven’t panned out in 35 years of Muni Metro service.
While not all of this has been bad (the Embarcadero turnaround is particularly useful, for example), we’ve spent hundreds of millions of dollars trying to make the existing Muni Metro work—and yet we still have trains backed up every single day.
We keep trying to pile an erratic series of streetcars through a single tunnel. It simply can’t handle even the artificially low frequency we run at today. In August, one out of every three days saw major systemic delays as a direct result of merging trains into and out of the tunnels.
There’s a moment when you have to evaluate whether your continual frantic obsession with symptoms ignores the underlying systemic problem.
Victims of our History
For well over a century, streetcars have been running through downtown San Francisco. Even at the turn of the 19th century, an impressively busy Market Street was one of the nation’s biggest transit corridors.
Muni keeps history alive by running what amounts to a streetcar system with an added tunnel under Market Street (somewhat of a happy accident correlated with the creation of BART.) While the tunnel provides a number of benefits over street-level transit, we fail to take advantage of the true opportunity it offers by still thinking about our transit network as a vestige of a system rooted in the early 1900’s.
I’m a huge believer in respecting our history and building on it, but we shouldn’t be running a scaled-down version of a century-old streetcar network in 2016; we should have something far better: a true metro network.
What is a “Metro” anyway?
Here’s the generally accepted definition of Metro from the UITP:
[Transit] operated on their own right of way and segregated from general road and pedestrian traffic. They are consequently designed for operations in tunnel, viaducts or on surface level but with physical separation in such a way that inadvertent access is not possible.
No Muni Metro line fits that definition (not even the new Central Subway.)
It’s time to draw the distinction between Muni Metro and Muni Tram — and expand both. Muni Tram would follow this UITP definition:
[Transit] not segregated from general road and pedestrian traffic, which share their right of way with general road and/or pedestrian traffic and are therefore embedded in their relevant national road traffic legislation (highway codes and specific adaptations).
(We’ve got this one down.)
So how do we make the Muni Metro…a Metro?
A few years ago, at Nextransit, we introduced our vision for M-Market, based off of studying 6 months of Muni train data, minute-by-minute, and building a full simulation to study alternatives.
The M-Market would be a true urban metro, meeting the aforementioned definition of a metro and running, consistently at 2–3 minute intervals, platform-length trains from Embarcadero to West Portal in 12 minutes flat—at all hours of the day.
What are the advantages of this underground over what we have today?
- Consistency. A train every 2–3 minutes means no need to check the weird TV underground only to discover you’re stuck waiting for 20 minutes.
- Reliability. By removing the unpredictability of street-running trams merging from above ground, we maintain a 100% automated subway.
- High Capacity. With platform-length trains, we get 430% of the capacity of a J, K, or T train (1-car), every single time.
- Faster Boarding, Less Time Stopping. When everyone boards every train, there’s no chaotic weaving through crowds on narrow platforms. Longer trains means less running down the platform or holding doors. Subway-optimized vehicles could even bring a 50% increase in doors, which drastically improves loading time.
Okay, but what happens above ground?
There are about 2,000 train entrances into the Market Street tunnel each day (and a similar number of exits). The data shows 1–4 minutes are spent per train making that transition, with a median just over 2 minutes. That’s about 150 hours of time our trains spend in limbo, rather than driving, or an additional 250–300 end-to-end train trips per day. Combined with almost twice that amount of time lost to unwanted delays and backups underground, we get 700–800 additional trips per day, no new vehicles required (though much desired!)
That translates to Muni Tram service running every 3–6 minutes above ground, with frequencies varied to match ridership needs on each line. This gives us a number of benefits above ground:
- Frequency. Muni Tram service can increase to every 3–6 minutes, depending on line, drastically reducing wait times. If fact, we can continue to increase the number of trains without impacting the subway (a major issue today.)
- Capacity. Increased frequency also translates directly to increased capacity, allowing us to get 50–100% more capacity per line.
- Reliability. The majority of delays have root causes related to subway backups and transitions. Our models show a 60–80% reduction in delays above ground.
- Ready for Expansion. Extend the N-Judah east along Duboce to SoMa. Extend the J-Church north along Fillmore. Extend the L east along Ocean to Bayshore Caltrain. Once you eliminate the requirement that trains enter the tunnel, the whole city is your oyster.
So how do we get to this real Metro/Tram thing?
We can technically make the decision to do this tomorrow, though the transfer points would be more of a pain-point than ideal. (Though is that really worse than consistent train backups?)
In fact, last month when the switches failed and trains couldn’t merge underground, Muni amusingly accidentally stumbled upon the solution, pointed out by @mazameli:
But, really, we should do this right. Here’s what we need to do to make this happen quickly.
First, we need to update our train car order today to include 47-foot car open-gangway six-car subway trains in the second delivery batch. This would exactly fit West Portal and Forest Hill Platforms, providing a 430% increase in capacity over a J, K, or T train today. (This could also easily scale up in the future to nine cars, the exact length of Church, Castro, and Van Ness.) M-Market doesn’t hinge on this, but it’s truly a good idea to do ASAP.
In the meantime, we can redesign the intersection of West Portal and Ulloa into West Portal Plaza, simplifying the transfer between the above-ground Tram and the Metro. (Eventually, West Portal Station could move entirely underground, in conjunction with the 19th Avenue Metro extension project.)
At the same time, Church Station entrances should be brought directly to the new N-Judah and J-Church platforms.
Finally, we tweak the E-Embarcadero route to share Market Street with the F-Market, allowing for a simple transfer at Embarcadero Station for travel along the Embarcadero in either direction.
Immediately, the entire system can breathe—now we can finally truly focus on optimizing each line independently for its own needs.
Moving Beyond
San Francisco needs a truly robust, modern transit network, and that’s at the core of how we must radically rethink mobility in San Francisco. But we can’t get there until we step back and re-evaluate what we have, and our current Muni Metro model simply isn’t built to scale—it can’t even keep up with today’s needs.
Our international peers are building hundreds of miles of new subways at a fraction of the cost, even in sandy, earthquake-prone densely populated cities. Paris alone is building 120 miles of new subway—the equivalent of 17 new cross-town subway lines in San Francisco—at a cost of $200 million/mile. Milan did it for half that price. (We’re running about 8x higher.)
I’d love to see our rail network grow in San Francisco, but we must first start by fixing what we have. If we can create a real Muni Metro, I’m confident we can build on that momentum to expand the network across the city.
If you’re interested in helping make this a reality, or just want to understand the concept better, read more at http://newmunimetro.com/ and sign up for the mailing list. You can also reach out to @nextransit or me @reedm with thoughts and ideas. It’s really time to make a coalition and turn this into a reality. When I look back at everything that’s been written about Muni Metro’s issues over the past 30 years, I really hope I’m not still writing the same thing in another decade.
